Complex xtUML models often contain many features that can be assembled from a set of simpler modes and behaviors. For these models, the xtUML process recommends that each mode or behavior be implemented using a short analysis, design, and test cycle. The products of these micro-cycles are then incrementally merged to form the more complex model. An important benefit of this approach is the marked improvement in model quality due to the repeated validation that is performed as new functionality is added. Yet for this approach to be productive, a test environment that is easily extended and supports sequence reuse is essential. This topic explores testbench structures that support these goals and demonstrate their value.